Image analysis for water penetration through clay brick veneer and stucco

Hi,

I am currently doing research on water penetration through clay brick veneer and stucco. The test runs for 4 hours and I capture images of penetrated water every 5 sec (Just to get more data). But during ImageJ analysis, I used images every 30 sec.
I find myself trouble during thresholding, as I want to track penetrated water on the face of the wall. Can I eliminate and select only the pattern I want other than thresholding and apply it to a

sequence of images?

Kindly help me out…

I have uploaded a couple of test images and I want to track penetrated water and draw a graph of damp area.

Thanks
Kshitij Ghanate
ghanate@uwindsor.ca

1 Like

Hello Mr. Ghanate,
It looks like you are set up perfectly to use the images as a stack. Then pick out the pattern/section the you want to analyze with the rectangle tool. Then Edit> Crop, the area of interest and analyze everything you want to know.
If you feel I misunderstood your question, then please reply and I will try to be more helpful.
Good Luck,
Bob

While performing/ adjusting the threshold, I am facing a lot of noise. In the above-uploaded images, I only want the threshold option to capture the dampness/ penetrated water. Auto/manual thresholding is capturing other grey bits on an 8-bit file, which I don’t want. Can somebody please help me in this situation?

Even if I crop, I am still facing noise during thresholding (manual/auto).

File size of your images is 2.5 MB.
Size of the files loaded in ImageJ is 68 MB (5184 x 3456 pixel).
The image is a highly compressed jpeg and contains strong jpeg artefacts.

If you want to reduce the ‘noise’ in your measurement my first hint would be to reduce the compression setting in your camera - or better use a file format with lossless compression.

If the following looks interesting to you … let me know.

2 Likes

Yes, it looks better from here. Can you please share the procedure?

The following code is just an example and not the only solution. Maybe it can give you some hints.
Open one image and execute the following code:

makeRectangle(1064, 1056, 2876, 2336);
run("Crop");
run("RGB Stack");
run("Gaussian Blur...", "sigma=4 stack");
run("Stack to Images");
imageCalculator("Multiply create 32-bit", "Green","Blue");
selectWindow("Result of Green");
rename("GxB");
run("Log");
run("Enhance Contrast", "saturated=0.35");
setOption("ScaleConversions", true);
run("8-bit");
setAutoThreshold("IsoData");
setOption("BlackBackground", false);
run("Convert to Mask");
rename(title + "_MASK");
selectWindow("Red");
close();
selectWindow("Green");
close();
selectWindow("Blue");
close();
2 Likes

@phaub
This succession:

is for me a discovery.
Thank you

3 Likes